VISIBLE AUTUMN MIGRATION AT ST. DAVID'S HEAD by W. E. WATERS INTRODUCTION St. David's Head (Pembrokeshire) lies at the south-western extremity of Cardigan Bay and appears geographically well placed for the study of visible bird migration. Published observations give few details of bird movement in this area. Lockley, Ingram and Salmon (1949) mention spectacular migration at several peninsulas, including St. David's Head, on the North Pembroke- shire coast. Snow (1953), in a general review of visible migration, gives directions of movement in this area for several species. The few other published references to St. David's Head add little more information. During the autumns of 1963 and 1964, I visited St. David's Head on 19 days (the year of observation is not repeated each time but can be found by referring to Table 3). Similar bird movements were seen during further watches in September 1966. In this paper, the Head refers to the tip of the peninsula only the headland refers to the whole of the area covered in this study which extends inland from the Head for about a mile to include the Coastguard station and the Porth Melgan area. From the easterly base of this triangular area all migrants flying to the north-west of Carn Llidi (595 feet) could be seen. However it was not possible for one observer to watch this and also the emigration and sea- passage best seen from the Head. My wife accompanied me on several visits but the limitations of solo observations became readily apparent. Visible migration studies in such an area would be enhanced by several observers with synchronous watches observing from different points. This paper deals only with visible migration, that is, migration actually in progress. Records of other migrants grounded on the headland during this study have already been published in Nature in Wales Field Notes. All time is G.M.T. SEA-WATCHES During the two autumns a total of 39 hours was spent in sea-watches looking north-west from a position on the Head about 100 feet above the sea. Watches were divided into 30 minute periods though several watches frequently followed successively, giving continuous observations of up to 3 hours at a time. All sea-watches were made with 7 X 50 binoculars and momentary rests from using these probably represent less than a tenth of the time spent sea-watching. The number of birds thus missed was reduced by quick scanning immediately before and after each brief rest.